Article

pubs.acs.org/cm

CH Bond Activation by Pd-substituted CeO2: Substituted Ions

versus Reduced Species
Lauren M. Misch, Joshua A. Kurzman, Alan R. Derk, Young-Il Kim, Ram Seshadri,*,, Horia Metiu,
Eric W. McFarland, and Galen D. Stucky,

Department of Chemistry & Biochemistry, Department of Chemical Engineering, and Materials Department and Materials
Research Laboratory, University of California, Santa Barbara, California 93106, United States

Department of Chemistry, Yeungnam University, Gyeongsan, Gyeongbuk 712-749, Republic of Korea

*
S Supporting Information

ABSTRACT: Substituted metal oxides containing ionic species have been

attracting a great deal of attention because of their potential ability to reduce
the usage of precious metals in heterogeneous catalysts. We investigate Pd-
substituted CeO2 for CH bond activation reactions including the partial
oxidation and dry reforming of CH4. This catalyst has been previously studied
for CO oxidation, NOx reduction, and the water-gas shift reaction. Pd-
substituted CeO2, Ce1xPdxO2, was prepared as a powder with high surface
area and a hollow sphere morphology using ultrasonic spray pyrolysis. The
catalysts were extensively characterized using synchrotron X-ray diffraction and
other techniques, confirming phase pure samples up to 10 mol % Pd
substitution. Ce0.95Pd0.05O2 was found to be active for partial oxidation of CH4 around 500 C and higher. Our studies,
including postcatalytic synchrotron diffraction, suggest that the single-phase Ce 1xPdxO2 material is not the active species and
that catalysis occurs instead over the reduced two-phase Pd 0/CeO2. This observation has been further confirmed by verifying the
activity of the reduced Pd0/CeO2 catalysts for ethylene hydrogenation, a reaction that is known to require Pd 0.
KEYWORDS: catalyst, nanoparticle, hydrocarbon

INTRODUCTION
Heterogeneous catalysis on the surfaces of platinum group
metals (PGMs) has long been studied, and the fundamental
processes are now understood in extraordinary detail from both
experimental and theoretical bases. Much less is known about
the surface chemistry of PGM species when they exist as ions in
solid-state materials such as simple and complex oxides. 1
In a recent review, Thomas2 promotes the idea of catalysis on
single active sites, well-separated from one another, in a manner
that mimics homogeneous catalysis. This important design
principle for novel approaches to element-efficient heteroge-
neous catalysis has been demonstrated through the use of ionic
species substituted on cation sites in metal oxides, including the
use of Pd2+-substituted metal oxides. Hegde and co-workers3
have shown conclusively that substituted PGM ions are active
for CO removal from gas streams,4 and as automotive three-
way catalysts.5 Pfefferle and co-workers6 have found that in
CH4 combustion, the role of PdO and related species is crucial.
It has been further suggested that the oxidized state of Pd 2+ (as
opposed to metallic Pd0) may be the most active species when
using substituted complex oxides for catalysis.79 Perhaps the
most compelling evidence for using Pd-substituted oxides for
heterogeneous catalysis comes from the successful application
of intelligent catalysts for automotive emissions control.10
In this contribution, we use ultrasonic spray pyrolysis (USP)
as a simple and clean method to prepare Pd-substituted CeO 2
catalysts with particle sizes in the sub-10 nm range. We have
characterized these materials using electron microscopy, X-ray
photoelectron spectroscopy, and synchrotron X-ray diffraction
(XRD) and observe that Pd substitutes in the lattice at least up
to x = 0.10 in Ce1xPdxO2.
In this work, Ce1xPdxO2 was tested for CH bond
activation reactions. While there are reports of complete and
partial CH4 oxidation over Pd,1115 a wealth of literature has
been published on Pd and CeO2 containing catalysts for CH4
combustion,1623 CO oxidation,2427 and NOx abatement.28
The oxygen storage capacity of CeO2 is enhanced when
substituted with Pd,2933 and doping promotes the formation
of oxygen vacancies in catalysis.5,34,35 It has also been
demonstrated that CeO2 supported catalysts and Pd supported
on CeO2 are active for the water-gas shift reaction.3638
Additionally, it was also shown that a Pd-containing catalyst
effectively converted CH 4 to a methanol derivative in
solution.39
The study of CH4 activation reactions has important energy
applications. While large reserves of CH4 exist and considerable
portions of these reserves are currently used to heat homes and
generate hydrogen for other synthetic processes, it is widely
Received: September 11, 2011
Revised: November 10, 2011
Published: November 16, 2011

2011 American Chemical Society 5432 dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439
Chemistry of Materials
accepted that the conversion of CH4 to liquid hydrocarbon
fuels efficiently with an inexpensive and robust catalyst would
be a substantial contribution to alternative energy research. 40
The usual FischerTropsch strategy requires oxidation to
mixtures of CO and H2, which are then converted to higher
hydrocarbons. Alternate partial oxidation strategies could be a
more direct route to valuable products. For example, dry
reforming of CH4 is of value because recovered CH4 is often
found in the presence of CO2. As large-scale separations are
expensive, it would be convenient to identify a catalyst that
efficiently converts CH4 to useful products in the presence of
CO2.41
We address the following questions in this work: (i) Does
USP provide a useful route to single-phase Pd-substituted CeO 2
with high surface area? (ii) Can Pd-substituted CeO2 be used as
a catalyst for CH4 activation in the presence of O2 (partial
oxidation) or CO2 (dry reforming)? (iii) Can it be concluded
that a substituted PGM ion is active for CH bond activation?
We found that Pd-substituted CeO2 behaves in a manner that is
nearly indistinguishable from supported Pd on CeO2 as a result
of the reduction of Pd2+ ions to Pd nanoparticles on CeO2. This
study complements prior work on the use of Pt substituted
CeO2 as a catalyst for CH4 activation.42
EXPERIMENTAL SECTION
Pd-substituted CeO2 was prepared using USP. The USP setup is based
on the apparatus described by Skrabalak et al.,43 which they used for
the preparation of nanoporous carbon. The precursor solution,
containing Ce(NO3)36H2O (99%, Aldrich) and Pd(NO3)22H2O
(99.999%, Aldrich) dissolved in the appropriate molar ratios in
Millipore water, was nebulized in the custom reaction vessel over a
Sunpenton humidifier. The precursor mist was carried by compressed
air through a vitreous silica tube in a Lindberg Blue/M tube furnace at
500 C. Product powders were collected in bubblers containing 4:1
H2O/EtOH. The suspensions were evaporated in crystallization dishes
at 80 C overnight, and the dry powder was collected.
Room temperature XRD data was collected on a Philips XPERT
diffractometer, and in situ variable temperature diffraction experiments
were performed with a Bruker D8 diffractometer equipped with an
Anton Parr hot-stage. Synchrotron X-ray powder diffraction patterns
were collected in transmission mode at room temperature on beamline
11-BM at the Advanced Photon Source, Argonne National Laboratory,
with an X-ray energy of near 30 keV. No evidence for sample
degradation or damage was observed. Rietveld refinements were
performed using the XND code.44 X-ray photoelectron spectra were
obtained on a Kratos Axis Ultra Spectrometer with a monochromatic
AlK source (E = 1486.7 eV). Samples were mounted on a stainless
steel sample holder using double-sided carbon tabs. The residual
pressure inside the analysis chamber was below 7 109 Torr. Survey
spectra over wide ranges of binding energy were acquired using an
analyzer pass energy of 160 eV, and spectra of Pd 3d levels were
acquired at a pass energy of 80 eV. Spectra were calibrated to the C 1s
peak from adventitious hydrocarbons, expected at a binding energy of
285.0 eV. For peak fitting of the spinorbit doublets in high resolution
scans, the d3/2 to d5/2 peak area was constrained to a ratio of 2/3.
Scanning electron micrographs (SEMs) were acquired on an FEI
XL40 Sirion FEG digital scanning microscope. SEM sample stages
were sputtered with Au plasma prior to imaging to reduce sample
charging. Transmission electron micrographs (TEMs) were taken on
an FEI Tecnai G2 Sphera Microscope. TEM copper-coated Cu grids
were prepared by dropcasting a dilute suspension of product in
ethanol onto grids. BrunauerEmmettTeller (BET) surface area
measurements were made on a MicroMetrics TriStar 3000
porosimeter using N2 as probe gas.
Catalytic testing was carried out in a home-built packed bed reactor,
equipped with MKS mass flow controllers and mass spectrometer
(SRS) for data acquisition. Quartz tubes (inner-diameter = 4 mm)
5433
Article
were packed with 25 mg of catalyst and 50 mg of HPLC grade
aluminum oxide (Aldrich) to prevent hotspots, with quartz wool plugs
on both ends of the powder. The loosely packed powder occupies a
length of 1 cm to maintain a space-time of 0.18 s with a total flow rate
of 30 sccm. Reactions were ramped from room temperature to 600 C
at a rate of 10 C/min. Catalysts were pretreated with Ar, 20% H 2/Ar
balance, or 20% O2/Ar balance. During partial oxidation of CH4 a 2:1
ratio of CH4/O2 was set to flow over the catalyst. This ratio is the
stiochiometric amount to produce synthesis gas: CH4 + 1/2 O2 CO
+ 2H2. All gases had a stated purity of better than 99.99%.
Anticipated products for partial oxidation reactions (2:1 CH 4/O2)
include combustion productions (CO2 and H2O), synthesis gas
(various ratios of H2 and CO), oxidative coupling products (C2H6,
C2H4, C2H2), and methanol. The mass spectrometer was set to record
the activity of the m/z ratios corresponding to these products. While
heating any hydrocarbon in the presence of oxygen to high
temperatures, combustion products are expected. It is possible for
unreacted CH4 to react with any CO2 produced from combustion and
proceed to do dry reforming of CH4 to produce synthesis gas (CO +
H2). If CO and H2 were produced during partial oxidation of CH4,
catalysts were tested for dry reforming of CH4, in which CH4 reacts
with CO2 produced from combustion. If CH4 reacts with CO2 to
produce synthesis gas, then this is a possible mechanistic route for
syngas production during partial oxidation reactions.
The anticipated products for dry reforming of CH4 (1:1 CH4/CO2)
are synthesis gas (H2 and CO). The mass spectrometer was set to
record the m/z ratios corresponding to all of the products listed for
partial oxidation reactions. All subsequent reactions were chosen based
on the materials behavior during partial oxidation conditions. Other
reactions studied in this work include ethylene hydrogenation (1:1
C2H4/H2) to 300 C heated at 10 C/min. Anticipated products of
this reaction include ethane and water. The results of characterization
of Ce1xPdxO2 and catalytic testing follow.
RESULTS AND DISCUSSION
Characterization of Ce1xPdxO2. Pd-substituted CeO2
(Ce1xPdxO2) was prepared with x = 0.025, 0.05, 0.075, 0.1
via USP. Representative scanning electron micrographs for a
sample with x = 0.05 of the as-prepared powders are shown in
Figure 1a. The hollow sphere morphology of the powders is
evident in the higher magnification image presented in the
inset. The morphology presumably results from evaporation of
liquid as the mist traveled through the furnace, leaving behind
polydispersed hollow spheres. Despite the relatively large size
of the agglomerates, the crystallites of which the spheres are
composed are rather fine, with grain sizes on the order of 5 nm,
as seen in the transmission electron micrographs of Figure 1b.
The Ce0.95Pd0.05O2 powder prepared by USP has BET surface
area of about 32 m2/g.
High-resolution synchrotron X-ray powder diffraction
patterns and corresponding Rietveld refinements are shown
in Figure 2 for the as-prepared samples and after calcination at
700 C for 16 h. Pure fluorite CeO2 is the only phase observed,
and no phase segregation occurs in any of the compositions.
We did not attempt to prepare materials with Pd concen-
trations higher than 10 mol %. The diffraction profiles are
broader in the substituted materials, relative to pure CeO 2, and
this is especially pronounced in the calcined samples (see inset
of Supporting Information, Figure S1). Refinements were
performed with models fixed at the nominal stoichiometry of
each sample ( = x) with Pd residing on the Ce site, and the
atomic displacement parameters (ADP) of Pd and Ce were
constrained to the same value. Because the Pd and O
occupancies, ADPs, and global scale factor are strongly
correlated, the occupancies cannot be refined. For this reason,
it is not possible to directly demonstrate the solid solubility of
dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439
Chemistry of Materials
Figure 1. (a) Scanning electron micrograph of Ce 0.95Pd0.05O2
particles prepared by nebulized USP showing a hollow sphere
morphology. (b) Transmission electron micrograph of the
Ce0.95Pd0.05O2 particles showing that hollow spheres are composed
of approximately 5 nm crystallites, confirming the correlation length
obtained from line broadening from synchrotron XRD.
Pd in CeO2 from average structure (Rietveld) refinement
techniques using XRD data.
Thermodiffraction data is presented in the Supporting
Information, Figure S1, showing the evolution of the 111 and
200 reflections of CeO2 between room temperature and 700 C
for unsubstituted CeO2 (x = 0) and 5% Pd-substituted CeO2 (x
= 0.05). The thermogravimetric analysis of 5% Pd-substituted
CeO2 (x = 0.05) is also shown in the Supporting Information,
Figure S2. Diffraction profiles of pure CeO2 narrow with
increasing temperatures, while the peak widths of the
substituted material remain significantly broader. The distinct
behaviors are indirect evidence that Pd is substituted into the
CeO2 lattice, and suggests that sintering may be suppressed in
the substituted material. These points are discussed in greater
5434
Article
Figure 2. (a) Synchrotron XRD data for as-prepared Ce1xPdxO2
with x = 0, 0.05, and 0.1. All samples are single-phase fluorite. Vertical
bars in the topmost panel indicate expected fluorite CeO 2 reflection
positions. (b) Diffraction from samples after calcining in air for 16 h at
700 C, with significant peak narrowing because of sintering. The inset
shows the strongest reflection, with height normalized, for samples
with increasing Pd2+ substitution, x. Samples with higher x values are
seen to possess significantly broader peaks.
detail with respect to analysis of synchrotron X-ray powder
diffraction studies of the Ce1xPdxO2 series.
There are considerable challenges associated with structure
determination and refinement of nanoscale materials from
Bragg scattering-based diffraction analysis. While it is known
that systematic errors arise in the determination of lattice
parameters for nanocrystalline materials,45 this is infrequently
acknowledged. Using the Debye function, Palosz and co-
workers simulated diffraction patterns for perfect SiC nano-
crystallites with sizes ranging from 3 to 8 nm and refined the
simulated data by the Rietveld method.45 Interestingly, they
found that in the approximation of a perfect experiment, that is,
no sample-offset error, the refined lattice constant was
systematically overestimated, increasingly as the crystallite size
decreased. In refinements approximating an imperfect experi-
ment in which the sample-offset error was allowed to float, the
refined lattice constant was increasingly underestimated as the
crystallite size decreased. Thus, even within the approximation
that a nanomaterial is a small single-crystalline piece of the bulk
material, Rietveld refinement fails to accurately extract the
lattice parameters. In this light, it is clear that great care must be
taken when establishing trends in the variation of lattice
parameters determined by Rietveld analysis.
The situation is further complicated by the fact that
nanocrystallites are not simply small portions of a bulk material.
Conventional crystallographic analysis operates on the
assumption that the environment of each lattice point is
identical. While this may be well approximated by atoms within
the core of a nanoparticle, it certainly does not apply to the
under-coordinated atoms at or near the surface. For this reason,
a single group of lattice constants does not capture the
complexity inherent to real nanocrystals. Palosz et al. have
dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439
Chemistry of Materials
extensively discussed limitations of Rietveld analysis for
structure determination in nanocrystalline materials.4547
With these limitations in mind, we address the observed
variation of the lattice constant as a function of substitution
level with caution. It is clear from the thermodiffraction and
synchrotron studies that Pd-substitution in CeO2 reduces the
XRD-coherent correlation length. In Figure 3a, the refined
Figure 3. (a) Variation of the cubic cell parameter of Ce1xPdxO2 as
a function of nominal Pd substitution x. Data are displayed for as-
prepared and calcined samples separately, as described in the text. The
cubic cell parameter was determined for the whole Q range of data,
and separately for data with Q > 8 1, shown with shaded symbols.
(b) Cell parameters displayed as a function of the reciprocal crystalline
correlation length as obtained from WilliamsonHall analysis of
synchrotron XRD data.
lattice parameters of the as-prepared and calcined samples are
plotted against the nominal Pd content x. One method for
reducing the error associated with lattice parameter determi-
nation in nanoparticles is to refine only the high Q portion of a
diffraction pattern, although this is only effective in the
approximation of a perfect crystallite. Despite this known
limitation, the lattice parameters from refining over the entire Q
range (open symbols) and only the high Q portion of the
patterns (shaded symbols, Q > 8 1) are compared in Figure
3a. It is immediately clear that the as-prepared samples appear
to have larger lattice constants than the calcined materials. The
lattice constants of the as-prepared materials are reduced when
only the high Q portions of the patterns are refined; the effect is
less pronounced in the calcined materials. Ce1xPdxO2 seems
to exhibit Vegard style behavior with a lengthening of a as the
Pd concentration is increased. This would be consistent with
increased cationcation repulsion arising from the removal of
oxygen because of the aliovalent substitution of Pd2+ for Ce4+.
However, the ShannonPrewitt ionic radius of 4-coordinate
Pd2+ (0.64 ) is significantly smaller than the radius of 8-
coordinate Ce4+ (0.97 ), so it is difficult to know whether the
observed expansion is an artifact of differences in crystalline
correlation lengths, or accurately representative of differences in
the lattice constants.
To address whether the variation in the lattice parameter
across the series is due to systematic differences in the
crystallite sizes, we performed a WilliamsonHall analysis48 on
5435
Article
each of the patterns. It is important to point out that neither
Scherrer nor WilliamsonHall analyses are quantitatively
accurate methods for extracting correlation lengths, though
they do provide reasonable first-order estimates. 45 Additionally,
in the case of nanocrystallites of relatively small sizes, < 10 nm,
the strain parameter extracted by the WH method carries little
physical meaning.47 We elected to do a WilliamsonHall
analysis because it involves fitting over the entire observed Q
range, but we note that similar estimates of the correlation
lengths were obtained by Scherrer analysis of a single reflection.
The refined lattice constants are plotted in Figure 3b as a
function of the inverse correlation length estimated from the
WH analysis; the volume weighted particle sizes and strains are
given in Table 1. As a function of the inverse correlation length,
Table 1. Volume Weighted Particle Sizes (Dv) and Strain ()
Obtained from a WilliamsonHall Analysis of Synchrotron
X-ray Powder Diffraction for Ce1xPdxO2 Where x = 0,
0.025, 0.05, 0.075, 0.1
calcined
as-prepared (00l) other reflections
x Dv () Dv () (%) Dv () (%)
0 72 15 293 2 0.03 1 296 9 0.06 3
0.025 64 11 157 4 0.13 1 149 5 0.16 1
0.05 60 5 118 3 0.11 1 137 1 0.27 1
0.075 53 7 123 5 0.11 2 134 6 0.24 1
0.10 55 12 101 0.16 102 5 0.30 2
the lattice parameters follow an approximately linear trend. It is
not possible to conclude whether the observed differences
result from systematic errors inherent to the Rietveld method,
or whether they genuinely reflect differences between the
samples. Within the associated error, the estimated strain is
almost constant across the series of as-prepared samples. In the
calcined materials, the estimated strain increases significantly in
going from the x = 0 unsubstituted material to x = 0.025, and
then gradually increases with the Pd concentration.
It is clear that Rietveld analysis cannot provide direct
evidence that Pd substitution occurs in CeO2, complicated by
the many factors we have discussed. Nonetheless, the fact that
these are single-phase materials displaying significantly different
behavior upon calcination compared to pure CeO2, coupled
with the observation that the XRD-coherent correlation length
changes as a function of Pd concentration, are highly suggestive
that Pd is dispersed in the CeO2 lattice. Verification that the
nominal Pd concentrations are indeed reflective of the actual
compositions is obtained by fully reducing the samples to two-
phase mixtures of CeO2 and fcc-Pd metal. A synchrotron XRD
pattern and corresponding Rietveld refinement for one such
sample (x = 0.05) are shown in Figure 4. The fcc-Pd
contribution is estimated to be 4.7 mol % by quantitative
phase analysis, in excellent agreement with the presumed Pd
content.
Scanlon et al. recently described an ab initio study of Pd and
Pt substitution in CeO2 and demonstrated that because of
crystal field stabilization effects, the PGM substitutents prefer
to displace off the ideal Ce lattice position by about 1.2 to
adopt square planar coordination, the most common
coordination geometry for d8 cations.33 This result emphasizes
the importance of applying structural probes that are sensitive
to local environments. While EXAFS studies have been
dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439
Chemistry of Materials
Figure 4. Synchrotron XRD data for Ce0.95Pd0.05O2 reduced in 5%
H2 in Ar at 700 C for 8 h. The Rietveld refinement shows fcc-Pd
metal to be quantitatively present with a mole ratio of 0.05.
reported on the Ce1xPdxO2 system,49,50 we are not aware of
any attempts to fit models similar to the one proposed by
Scanlon et al.
X-ray photoelectron spectroscopy (XPS) of the Pd 3d region
was investigated to determine the charge state of substituted Pd
in Ce0.95Pd0.05O2. Indicated in Figure 5 are the binding
Figure 5. X-ray photoelectron spectrum of the Pd 3d region of as-
prepared Ce0.95Pd0.05O2 acquired with a pass energy of 80 eV. The
Pd 3d5/2 peak is shifted to higher binding energy than found in PdO
(dashed line indicates position) suggesting a more ionic charge state
than that of PdO. No evidence for metallic Pd is seen.
energies for the Pd 3d5/2 signal in PdO (336.8 eV) and Pd
metal (335.4 eV).51 In Figure 5, the Pd 3d5/2 signal for
Ce0.95Pd0.05O2 is seen at 337.4 eV, a slightly higher binding
energy than that of PdO or Pd metal. The increased ionic
character suggests Pd lattice substitution. This shift to higher
binding energy is in agreement with the XPS of the Pd 3d
region taken by Singh et al. for Ce0.95Pd0.05O2 prepared via
solution combustion synthesis in which the 3d5/2 signal is seen
at 337.4 eV.52 Though this does not entirely rule out the
possibility of PdO clusters on the surface, both bulk probes like
diffraction and surface probes like XPS suggest ionic Pd is
incorporated into the CeO2 lattice. The purpose of this study
5436
Article
was also to determine if amorphous PdO, undetectably by
XRD, was present on the sample. Because of the low resolution
of the XPS data, it would not be possible to deconvolute the
two Pd2+ signals. However, if amorphous PdO were present, it
would crystallize upon calcination.
Reactivity Studies. The Ce1xPdxO2 series was tested
for CH bond activation in partial oxidation of CH4 and dry
reforming of CH4. Partial oxidation was tested over pure CeO2
as a control (Figure 6a). Even at 600 C there is no conversion
Figure 6. Partial oxidation of methane (2:1 CH4/O2) in Ar heated at
10 C/min to 600 C over (a) CeO2, (b) preoxidized (20% O2 in Ar
to 500 C 1 h) Ce0.95Pd0.05O2, (c) prereduced (20% H2 in Ar to 500
C 1 h) Ce0.95Pd0.05O2, and (d) prereduced (20% H2 in Ar to 500 C
1 h) PdO/CeO2. Almost no reaction is observed over pure CeO2. The
substituted Pd catalyst produces combustion products and non-
stoichiometric synthesis gas during partial oxidation and behaves
similarly to the supported Pd catalyst under reaction conditions.
of CH4 to products of interest. Whereas the USP prepared
CeO2 was inactive, Figure 6 shows the Pd-substituted CeO2
catalyst is active for CH4 combustion.
It was of interest to use a minimum amount of PGM while
still achieving CH bond activation. The quantitative work was
performed primarily on Ce0.95Pd0.05O2 which was found by us
to be slightly more active than Ce 0.975Pd0.025O2 and
approximately the same as Ce0.925Pd0.075O2. For partial
oxidation of CH4, Ce0.95Pd0.05O2 was subjected to either
prereduction or preoxidation. In both pretreatment cases, the
same reaction character is observed. The only difference the
pretreatment yields is a slightly lower activation temperature for
the prereduced sample. In partial oxidation over
Ce0.95Pd0.05O2 (Figures 6b,c), combustion products (CO2
and H2O) were observed along with nonstoichiometric
synthesis gas. Excess H2 is produced from the partial oxidation
of CH4 over Ce0.95Pd0.05O2.
To consider if CH4 is reacting with CO2 produced from
combustion, dry reforming of CH4 was tested separately. The
same two pretreatments were performed individually. The
pretreatment makes very little difference in the reaction
character and the activation temperature of the catalyst. Very
minimal synthesis gas was producing during dry reforming of
CH4 over Ce0.95Pd0.05O2. Dry reforming is not the sole
mechanism by which excess H2 is produced.
dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439
Chemistry of Materials
To further probe the mechanism by which partial oxidation
of CH4 over Ce0.95Pd0.05O2 produces excess H2 a steady state
reaction was performed. While the gas ratios for partial
oxidation remained the same, the temperature ramp was
changed to allow the catalyst to come to steady state at each
temperature stage before continuing. The temperature ramping
for the steady state reaction was 1 h dwells in 50 C increments
starting at 400 C, increasing to 600 C, and back down again.
As shown in Figure 7, the steady state reaction clearly shows
Figure 7. Steady state partial oxidation of methane (2:1 CH4/O2) in
Ar heated in 50 C increments from 450 to 600 C with a dwell time
of 1 h at each temperature step over Ce0.95Pd0.05O2. At 450 C
sufficient Pd metal is present to produce nonstoichiometric synthesis
gas, H2 in excess. After combustion, several secondary reactions occur
including steam reforming and water gas shift.
consumption of water during partial oxidation of CH 4,
suggesting steam reforming. Along with the consumption of
water, more CO2 is produced than expected. This is likely a
result of the water-gas shift reaction. The excess H2 observed
during partial oxidation of CH4 over Ce0.95Pd0.05O2 also likely
results from a combination of some dry reforming, steam
reforming, and water gas shift reactions. The long time steady
state reaction for partial oxidation of methane over
Ce0.95Pd0.05O2 was carried out for 24 h at 600 C (Supporting
Information, Figure S3).
To explain why prereduction causes this catalyst to become
active at a slightly lower temperature than the preoxidized
sample, we chose to further investigate how reducing
conditions affected this material. Synchrotron powder XRD
of the prereduced catalyst in Figure 4 distinctly shows fcc-Pd in
addition to cubic CeO2. It appears that Pd-substituted CeO2
becomes Pd supported on CeO2 under reducing conditions,
and this is the catalytically active phase for CH bond
activation. Pd-substituted CeO2 is not active for CH bond
activation. To confirm that the substituted material behaves like
Pd metal under reaction conditions, partial oxidation was
recorded for a prereduced sample of PdO supported on CeO 2
(Figure 6d). Partial oxidation of CH4 over Pd metal supported
on CeO2 shows identical reaction character to partial oxidation
over the Pd-substituted CeO2, but with activity for excess
hydrogen production igniting at about 400 C as opposed to
about 450 C in Pd-substituted CeO2.
5437
Article
The reaction character observed for partial oxidation and dry
reforming of methane over Pd-substituted CeO2 is in contrast
to that Pt-substituted CeO2. Partial oxidation of methane over
Pt-substituted CeO2 does produce stoichiometric synthesis gas
between 450 and 500 C, while Pd-substituted CeO2 produces
nonstoichiometric synthesis gas in the form of excess H2. Pt-
substituted CeO2 is also active for dry reforming of methane to
synthesis gas with relatively high conversion, while Pd-sub ceria
produces nonstoichiometric synthesis gas with a very low yield.
Pd-substituted CeO2 likely undergoes several secondary
reactions during partial oxidation of methane, including dry
reforming of methane, water-gas shift, and steam reforming of
methane
The synchrotron powder diffraction pattern was collected for
postreaction Ce0.95Pd0.05O2 mixed with Al2O3. Just as fcc-Pd
was seen in the reduced material (Figure 4), so too is this phase
observed in the diffraction pattern shown in Figure 8 along with
Figure 8. Synchrotron XRD data for Ce0.95Pd0.05O2 mixed with -
Al2O3 after use as catalyst under partial oxidation conditions to 600
C. The Rietveld refinement shows fcc-Pd metal present along with
diluent Al2O3. The asterisks indicate an unidentified impurity.
the -Al2O3 diluent. We took special care to cool the material in
an inert atmosphere after becoming active under reaction
conditions. The material was also handled carefully, quickly
contained, and promptly sent for characterization. We
recognize that some reoxidation may take place but the fcc-
Pd phase is very clearly seen in the postreaction material.
Certainly no PdO phase is observed in the postreaction
material. Moreover, the fact that the catalytic behavior matches
that of Pd/CeO2 further supports that the postreaction material
does contain Pd metal. The correlation length of fcc-Pd
determined from synchrotron XRD and the Scherrer line
broadening equation is near 7 nm for postreaction
Ce0.95Pd0.05O2 and near 100 nm for the as-prepared Pd/
CeO2 used for comparison. However, it should be noted that
the Scherrer line broadening equation does not provide the
most accurate measure of coorelation length at these length
scales. It seems that the catalytically active phase of
Ce0.95Pd0.05O2 for partial oxidation of CH4 is actually the
reduced Pd supported on CeO2. Other reactions and chemical
probes were considered to determine the presence of metallic
Pd in this catalyst.
dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439
Chemistry of Materials Article

It would appear that Pd-substituted CeO2 becomes Pd

supported on CeO2 under reaction conditions. Since Pd
supported on oxides is capable of catalyzing ethylene
hydrogenation, we performed ethylene hydrogenation over
Ce0.95Pd0.05O2.53 As seen in Figure 9a, no ethane was

Figure 10. Ethylene hydrogenation (1:1 C2H4/H2) over as-prepared

Ce0.95Pd0.05O2 heated at 10 C/min to 300 C. The material
becomes active and produces C2H6 only once the material become
sufficiently reduced around 120 C, further confirming that the
catalytically active form of this material is Pd supported on CeO 2.

the preparation of substituted metal oxides and potentially for

Figure 9. Ethylene hydrogenation (1:1 C2H4/H2) in Ar heated at 10
the preparation of well-dispersed metal nanoparticles on oxide
C/min to 300 C over (a) CeO2, (b) prereduced (20% H2 in Ar to supports upon reduction.
300 C 1 h) Ce0.95Pd0.05O2, (c) prereduced (20% H2 in Ar to 300 C
1 h) PdO/CeO2. The reduced Pd substituted catalyst is active for a
reaction known to take place on Pd metal and performs similarly to

*
ASSOCIATED CONTENT
S Supporting Information
the supported Pd catalyst.
Further details are given in Figures S1S3. This material is
produced over unsubstituted CeO2. However, ethane was available free of charge via the Internet at http://pubs.acs.org.
produced over prereduced Ce0.95Pd0.05O2 at room temper-
ature (Figure 9b). Increasing the temperature did not increase
the ethylene conversion in this reaction. The same behavior was
AUTHOR INFORMATION
Corresponding Author
observed for reduced PdO/CeO2 (Figure 9c). It is interesting *E-mail: seshadri@mrl.ucsb.edu.
to note that there appears to be a slightly higher selectivity
toward ethane production for reduced Ce0.95Pd0.05O2 than the
reduced PdO/CeO2 used for comparison. To confirm that this
reaction proceeds over Pd metal supported on CeO2, and not
ACKNOWLEDGMENTS
This work was supported by the Department of Energy, Office
of Basic Energy Sciences through Grant DE-FG02-89ER14048
the as-prepared Ce0.95Pd0.05O2, we attempted ethylene
and Grant DE-FG02-10ER16081. L.M.M. thanks Janet Kayfetz
hydrogenation without a prereduction. The catalyst does not
for suggestions on the manuscript. G.D.S. acknowledges
become active until it becomes sufficiently reduced by the
support from the University of California Discovery Grant
ethylene and hydrogen flowing over the catalyst. Figure 10
Program GCP08-128649. Use of the Advanced Photon Source
shows that at around 110 C the catalyst was reduced to Pd
was supported by the U.S. Department of Energy, Office of
metal supported on CeO2 at which point ethylene was
Science, Office of Basic Energy Sciences, under Contract No.
converted to ethane. The catalyst continued to actively produce
DE-AC02-06CH11357. L.M.M. is supported by a Fellowship
ethane while it was cooled back to room temperature.
from the ConvEne IGERT Program (NSF-DGE 0801627).
CONCLUSIONS
Pd-substituted CeO2 catalysts have been successfully prepared
A.R.D. acknowledges the NSF for a Graduate Research
Fellowship under Grant DGE 0707430. The use of the shared
facilities of the UCSB Materials Research Laboratory: an NSF
via USP with a surface area of 32 m2/g and hollow sphere MRSEC (DMR-1121053), a member of the NSF-funded
morphology. These materials are phase pure up to 10 mol % Pd Materials Research Facilities Network (www.mrfn.org) is
substitution. This material becomes catalytically active for CH gratefully acknowledged.
bond activation only after the Pt2+ ions have been reduced to
Pd metal supported on CeO2. Partial oxidation of CH4 over
Ce0.95Pd0.05O2 yields the expected combustion products along
with nonstiochiometric synthesis gas in the form of excess
REFERENCES
(1) Crabtree, R. H. The organometallic chemistry of the transition
metals; John Wiley & Sons Inc: New York, 2009; p 505.
hydrogen gas. The excess hydrogen is a result of several (2) Thomas, J. M. J. Chem. Phys. 2008, 128, 182502.
secondary reactions occurring after combustion, including dry (3) Hegde, M. S.; Madras, G.; Patil, K. C. Acc. Chem. Res. 2009, 42,
reforming of CH4, steam reforming of CH4, and water gas shift. 704712.
The catalytically active phase for this material is Pd supported (4) Bera, P.; Priolkar, K. R.; Gayen, A.; Sarode, P. R.; Hegde, M. S.;
on CeO2, confirmed by the ethylene hydrogenation reaction. Emura, S.; Kumashiro, R.; Jayaram, V.; Subbanna, G. N. Chem. Mater.
Additionally, we have identified USP as an adequate method for 2003, 15, 20492060.

5438 dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439

Chemistry of Materials Article

(5) Baidya, T.; Marimuthu, A.; Hegde, M. S.; Ravishankar, N.; (44) Berar, J.; Baldinozzi, G. Procedures for the refinement of
Madras, G. J. Phys. Chem. C 2007, 111, 830839. incommensurate structures using XND. Coding issues for the
(6) Datye, A.; Bravo, J.; Nelson, T.; Atanasova, P.; Lyubovsky, M.; refinement of incommensurate structures In Newsletter No. 5,
Pfefferle, L. Appl. Catal., A 2000, 198, 179196. Commission on Crystallographic Computing; Billinge, S., Chapuis, G.,
(7) Singh, U.; Li, J.; Bennett, J.; Rappe, A.; Seshadri, R.; Scott, S. J. Cranswick, L.. Lifshitz, R., Eds.; International Union of Crystallog-
Catal. 2007, 249, 349358. raphy, 2005; pp 59.
(8) Li, J.; Singh, U. G.; Bennett, J. W.; Page, K.; Weaver, J. C.; Zhang, (45) Palosz, B.; Grzanka, E.; Gierlotka, S.; Stel makh, S.; Pielaszek,
J.-P.; Proffen, T.; Rappe, A. M.; Scott, S.; Seshadri, R. Chem. Mater. R.; Bismayer, U.; Neuefeind, J.; Weber, H. P.; Proffen, T.; Dreele, Von,
2007, 19, 14181426. R.; Palosz, W. Z. Kristallogr. 2002, 217, 497509.
(9) Li, J.; Singh, U. G.; Schladt, T. D.; Stalick, J. K.; Scott, S. L.; (46) Palosz, B.; Grzanka, E.; Gierlotka, S.; Stelmakh, S.; Pielaszek, R.;
Seshadri, R. Chem. Mater. 2008, 20, 65676576. Lojkowski, W.; Bismayer, U.; Neuefeind, J.; Weber, H. P.; Palosz, W.
(10) Nishihata, Y.; Mizuki, J.; Akao, T.; Tanaka, H.; Uenishi, M.; GPHT 2003, 76, 171185.
Kimura, M.; Okamoto, T.; Hamada, N. Nature 2002, 418, 164167. (47) Palosz, B.; Grzanka, E.; Gierlotka, S.; Stelmakh, S. Z. Kristallogr.
(11) Ribeiro, F.; Chow, M. J. Catal. 1994, 146, 537544. 2010, 225, 588598.
(12) Escandon, L.; Ordonez, S.; Vega, A. Chemosphere 2005, 58, 9 (48) Williamson, G.; Hall, W. Acta Metall. 1953, 1, 2231.
17. (49) Baidya, T.; Priolkar, K. R.; Sarode, P. R.; Hegde, M. S.; Asakura,
(13) Hicks, R.; Qi, H.; Young, M. J. Catal. 1990, 122, 280294. K.; Tateno, G.; Koike, Y. J. Chem. Phys. 2008, 128, 124711.
(14) Basile, A.; Paturzo, L. Catal. Today 2001, 67, 6575. (50) Priolkar, K. R.; Bera, P.; Sarode, P. R.; Hegde, M. S.; Emura, S.;
(15) Cullis, C. F.; Nevell, T. G.; Trimm, D. L. J. Chem. Soc., Faraday Kumashiro, R.; Lalla, N. P. Chem. Mater. 2002, 14, 21202128.
Trans. 1 1972, 68, 1406. (51) Brun, M.; Berthet, A.; Bertolini, J. J. Electron Spectrosc. 1999,
(16) Muller, C.; Maciejewski, M.; Koeppel, R.; Baiker, A. J. Catal. 104, 5560.
1997, 166, 3643. (52) Singh, P.; Hegde, M. S. Cryst. Growth Des. 2010, 10, 2995
(17) Muller, C.; Maciejewski, M.; Koeppel, R.; Baiker, A. Catal. 3004.
Today 1999, 47, 245252. (53) McGown, W. T.; Kemball, C.; Whan, D. A.; Scurrell, M. S. J.
(18) Ciuparu, D.; Altman, E.; Pfefferle, L. J. Catal. 2001, 203, 6474. Chem. Soc., Faraday Trans. 1: 1977, 73, 632647.
(19) Groppi, G.; Cristiani, C.; Lietti, L.; Ramella, C.; Valentini, M.;
Forzatti, P. Catal. Today 1999, 50, 399412.
(20) Carstens, J.; Su, S.; Bell, A. J. Catal. 1998, 176, 136142.
(21) Strobel, R.; Grunwaldt, J.; Camenzind, A.; Pratsinis, S.; Baiker,
A. Catal. Lett. 2005, 104, 916.
(22) Grunwaldt, J.-D.; Maciejewski, M.; Baiker, A. Phys. Chem. Chem.
Phys. 2003, 5, 14811488.
(23) Persson, K.; Ersson, A.; Jansson, K.; Iverlund, N.; Jaras, S. J.
Catal. 2005, 231, 139150.
(24) Roy, S.; Hegde, M. S.; Ravishankar, N.; Madras, G. J. Phys.
Chem. C 2007, 111, 81538160.
(25) Roy, S.; Marimuthu, A.; Hegde, M. S.; Madras, G. Appl. Catal., B
2007, 73, 300310.
(26) Bera, P. J. Catal. 2000, 196, 293301.
(27) Baidya, T.; Dutta, G.; Hegde, M. S.; Waghmare, U. V. Dalton
Trans. 2008, 455.
(28) Roy, S.; Hegde, M. S.; Madras, G. Appl. Energy 2009, 86, 2283
2297.
(29) Singh, P.; Hegde, M. S. J. Solid State Chem. 2008, 181, 3248
3256.
(30) Gupta, A.; Waghmare, U. V.; Hegde, M. S. Chem. Mater. 2010,
22, 51845198.
(31) Baidya, T.; Gayen, A.; Hegde, M. S.; Ravishankar, N.; Dupont,
L. J. Phys. Chem. B 2006, 110, 52625272.
(32) Baidya, T.; Gupta, A.; Deshpandey, P. A.; Madras, G.; Hegde,
M. S. J. Phys. Chem. C 2009, 113, 40594068.
(33) Scanlon, D. O.; Morgan, B. J.; Watson, G. W. Phys. Chem. Chem.
Phys. 2011, 13, 4279.
(34) Haneda, M.; Mizushima, T.; Kakuta, N. J. Phys. Chem. B 1998,
102, 65796587.
(35) Ciuparu, D.; Bozon-Verduraz, F.; Pfefferle, L. J. Phys. Chem. B
2002, 106, 34343442.
(36) Fu, Q. Science 2003, 301, 935938.
(37) Gorte, R.; Zhao, S. Catal. Today 2005, 104, 1824.
(38) Sharma, S.; Deshpande, P. A.; Hegde, M. S.; Madras, G. Ind.
Eng. Chem. Res. 2009, 48, 65356543.
(39) Kao, L. C.; Hutson, A. C.; Sen, A. J. Am. Chem. Soc. 1991, 113,
700701.
(40) Lunsford, J. Catal. Today 2000, 63, 165174.
(41) Bradford, M.; Vannice, M. Catal. Rev. 1999, 41, 142.
(42) Tang, W.; Hu, Z.; Wang, M.; Stucky, G. D.; Metiu, H.;
McFarland, E. W. J. Catal. 2011, 273, 125137.
(43) Skrabalak, S. E.; Suslick, K. S. J. Am. Chem. Soc. 2006, 128,
1264212643.

5439 dx.doi.org/10.1021/cm202709y | Chem. Mater. 2011, 23, 54325439